URGE Pod of the Department of Earth and Atmospheric Sciences
In the spring of 2021, a group of EAS department members formed a ‘pod’ to participate in the NSF-funded Unlearning Racism in Geoscience (URGE) curriculum [https://urgeoscience.org/]. URGE organized a series of biweekly readings and expert interviews aimed at deepening our knowledge of the effects of racism on the participation and retention of Black, Brown, and Indigenous people in Geoscience. Pods from over 300 institutions around the world participated.
The pod produced a set of ‘deliverables’ in which we considered ways to make department policies more anti-racist; these can be found online [https://urgeoscience.org/pods/cornell-earth-and-atmospheric-sciences/]. The pod continues to work on how these ideas can be integrated into the policies and practices of the department, and welcomes all interested participants.
How to create good scientific figures
Principles of scientific figures:
- As self-explanatory as possible. Include panel titles, axes labels with units, legends within panels, annotations, etc. Inspirational goal: other people should be able to (and want to) use your figure unaltered in a talk of their own.
- Color schemes. Avoid ‘jet’-type color bars. Use divergent color bars for anomalies. Choose consistent colors throughout a study. Colorblind-friendly colors.
- Figure fonts should be sans serif. Figure font sizes should not be smaller than the smallest font in the typeset paper (often the caption).
- Make vector-based figures (.pdf, .eps, .svg) rather than raster-based figures (.jpg, .png). If you do make raster-based figures, make them high-resolution (150+ dpi).
Short guide to creating figures in python by graduate student Yan-Ning Kuo (May 2025).
How to create good scientific presentations
Some basic ideas that I think help:
- Think about your audience. What do they know already? What will they be most interested in?
- Show only 1-2 figures per slide. If you really must show more than 2 figures on a slide, build that slide step by step. Humans cannot digest 12 panels while listening to you talking about one of the panels.
- Make figures large enough. Axis labels should be readable from the back of the room. (Label your axes)
- Annotate figures in case you don’t have a laser pointers.
- Limit text. On a slide with figures, add at most one sentence on the key takeaway from the figure.
- Get to the point (first). Attention spans are short. But do introduce axes labels and color bars of complicated figures if you plan to spend time on them.
- Learn from others! If you see a memorable presentation, think why it worked. I’ve learned a lot from Scott St. George’s tips and of course Richard Alley’s way of explaining things.
Thoughts on AI
This is likely going to evolve but I’m trying to curate some readings and thoughts on the topic:
- Reverse Centaur’s Guide to AI. On the corporate motivation for AI deployment.
- AI for student learning. Seems currently limited. If they don’t learn the basics and critical thinking, they will struggle to determine whether AI is correct or not. It is like an EOF analysis of Z500. Everyone can calculate it but without the statistical and atmospheric science background, it is difficult to evaluate the results. Once you know how things work, AI can become a useful tool.